You are on page 1of 10

Increasing consumption of sugar-sweetened beverages among US

adults: 1988–1994 to 1999–20041–3


Sara N Bleich, Y Claire Wang, Youfa Wang, and Steven L Gortmaker

ABSTRACT research has examined these factors simultaneously. The dis-


Background: Consumption of sugar-sweetened beverages (SSBs) proportionate effect of obesity on minority and low socioeco-
has been linked to obesity and type 2 diabetes. nomic status (SES) communities (2, 3, 12–14), the positive
Objective: We examined national trends in SSB consumption association between intention to lose weight and weight control
among US adults by sociodemographic characteristics, body weight behaviors (15–17), the significant effect of even modest weight
status, and weight-loss intention. loss on the elimination or reduction of adverse health conditions
Design: We analyzed 24-h dietary recall data to estimate beverage associated with obesity (18), and evidence suggesting that cal-
consumption among adults (aged 20 y) obtained from the third ories in liquid may be less well compensated than calories from

Downloaded from ajcn.nutrition.org by guest on September 5, 2012


National Health and Nutrition Examination Survey (NHANES III, solid foods (19, 20) make this is an important area of study.
1988–1994; n ¼ 15979) and NHANES 1999–2004 (n ¼ 13431). Although the complex interactions between factors that cause
Results: From 1988–1994 to 1999–2004 on the survey day, the obesity are not fully understood, it is widely accepted that
percentage of adult SSB drinkers increased from 58% to 63% a positive energy balance (higher caloric intake than energy
(P , 0.001), per capita consumption of SSB increased by 46 expenditure) leads to weight gain (21, 22). Reduction of SSBs in
kcal/d (P ¼ 0.001), and daily SSB consumption among drinkers the adult diet may help prevent weight gain and promote weight
increased by 6 oz (P , 0.001). In both survey periods, per capita loss. Such efforts may offer relatively simple, low-cost solutions
SSB consumption was highest among young adults (231–289 kcal/d) to weight reduction and are consistent with recommended bev-
and lowest among the elderly (68–83 kcal/d). Young blacks had the erage consumption patterns (23, 24). Moreover, the identifica-
highest percentage of SSB drinkers and the highest per capita con- tion of variations in beverage consumption by subpopulation
sumption compared with white and Mexican American adults (P ,
groups will be useful for the development of targeted policies or
0.05). Overweight-obese adults with weight-loss intention (compared
nutrition programs aimed at reducing intake of ‘‘empty calo-
with those without) were significantly less likely to drink SSB, but
ries’’ among adults, particularly those from SSBs.
they still consumed a considerable amount in 1999–2004 (278 kcal/d).
The purpose of this study was to describe national changes
Among young adults, 20% of SSB calories were consumed at work.
(1988–1994 to 1999–2004) in the percentage of drinkers, amount
Conclusions: Over the past decade, US adult SSB consumption has
consumed, consumption location, and type of beverage among
increased. SSB comprises a considerable source of total daily intake
US adults by sociodemographic characteristics, weight category,
and is the largest source of beverage calories. SSB consumption is
highest among subgroups also at greatest risk of obesity and type and weight-loss intention. This analysis does not attempt to
2 diabetes. Am J Clin Nutr 2009;89:372–81. estimate the effect of SSB intake on obesity incidence, given our
reliance on cross-sectional data. Other research, supported by
longitudinal data, provides strong evidence for a causal relation
between SSB intake and increased body mass index (BMI; in kg/m2)
(1, 4, 25–27).

INTRODUCTION
1
From the Departments of Health Policy and Management (SNB), Center
Consumption of sugar-sweetened beverages (SSBs) has been
for Human Nutrition and International Health (YW), Johns Hopkins Bloom-
linked to the obesity epidemic (1), which currently affects one- berg School of Public Health, Baltimore, MD; the Department of Health
third of US adults (2, 3), and type 2 diabetes (4). From 1977 to Policy and Management, Columbia Mailman School of Public Health,
2001, energy intake from soft drinks and fruit drinks increased New York, NY (YCW); and the Department of Society, Human Develop-
by 135% (5), and the prevalence of adult obesity doubled (6). ment and Health, Harvard School of Public Health, Boston, MA (SLG).
2
Over the same period, the percentage of calories from all bev- Supported by the Department of Health Policy and Management at the
erages increased by .50% (7). Bloomberg School of Public Health and the Robert Wood Johnson Founda-
Much of the literature on adult beverage consumption has tion, grant 57891.
3
Address reprint requests and correspondence to SN Bleich, Department
focused on specific drink types (8, 9) or on broad temporal trends
of Health Policy and Management, Bloomberg School of Public Health,
and patterns (5, 7, 10). With the exception of a few studies (5, Johns Hopkins University, 624 North Broadway, Room 451, Baltimore,
11), little research has focused on national changes in adult MD 21205. E-mail: sbleich@jhsph.edu.
beverage consumption by sociodemographic groups, body Received August 26, 2008. Accepted for publication October 22, 2008.
weight status, and weight-loss intention. To our knowledge, no First published online December 3, 2008; doi: 10.3945/ajcn.2008.26883.

372 Am J Clin Nutr 2009;89:372–81. Printed in USA. Ó 2009 American Society for Nutrition

Supplemental Material can be found at:


http://ajcn.nutrition.org/content/suppl/2009/03/03/ajcn.2008.
26883.DC1.html
INCREASING CONSUMPTION OF SSBs AMONG US ADULTS 373
SUBJECTS AND METHODS Body weight status
Data and design In both surveys, weight and height were measured with the use
of standard procedures in a mobile examination center. Normal
Nationally representative data from the third National Health weight was defined as a BMI from 18.5 to 24.9; overweight was
and Nutrition Examination Survey (NHANES III; 1988–1994) and defined as BMI from 25 to 29.9, and obese was defined as BMI 
NHANES 1999–2004 were used. The NHANES is a population- 30 (28).
based survey designed to collect information on the health and
nutrition of the US population. Participants were selected based Intention to lose weight
on a multistage, clustered, probability sampling strategy. Survey
Respondent intention to lose weight was assessed by the
respondents are representative of the noninstitutionalized US
survey question, ‘‘During the past 12 months, have you tried to
population. Since 1999, data have been collected annually. Our
lose weight?’’ Respondents who answered ‘‘yes’’ were catego-
analysis combined the continuous NHANES data collection
rized as trying to lose weight.
(1999–2004) and compared it with NHANES III (1988–1994). A
complete description of data collection procedures and analytic
Socioeconomic status
guidelines are available elsewhere (www.cdc.gov/nchs/nhanes.htm).
The poverty:income ratio (PIR; the ratio of household income
to a family’s appropriate poverty threshold) was based on self-
reported household income. We dichotomized the PIR into lower
Study sample and higher income groups based on eligibility for food assistance
The study sample consists of adults aged 20 y with com- programs (ie,  130% of the poverty level). Education was

Downloaded from ajcn.nutrition.org by guest on September 5, 2012


pleted 24-h dietary recalls from the 2 cross-sectional NHANES categorized into the following mutually exclusive 3 categories: 1)
surveys. Survey respondents were excluded if they were pregnant less than high school, 2) high school (or GED), and 3) more than
at the time of data collection or if their dietary recall was in- high school.
complete or unreliable (as determined by the NHANES staff).
Because of the small sample size of the other race-ethnicity Analysis
category, we only included non-Hispanic white, non-Hispanic
black (hereafter referred to as whites and blacks), and Mexican All analyses were weighted to be representative of the general
Americans in the analyses. population and conducted using STATA, version 9.2 (Stata Corp,
College Station, TX) to account for the complex sampling
structure. Multivariate regressions were used to adjust for po-
tential changes in population composition over the 2 time
Measures periods, including race-ethnicity, sex, income, age, marital status,
Beverages employment status, and education. All tables and figures report
predicted means based on the adjusted models.
Survey respondents reported all food and beverages consumed
in a prior 24-h period (midnight to midnight) and reported type,
quantity, time, and location of each food and beverage con- RESULTS
sumption occasion. After the dietary interview, all reported food The 2 samples had comparable distributions of sex, age,
and beverage items were systemically coded with the use of the employment status, and income status, but the 1999–2004 sample
US Department of Agriculture (USDA) Food and Nutrient Da- had significantly fewer non-Hispanic whites and normal-weight
tabase. Caloric content and other nutrients derived from each persons, and significantly more Mexican Americans, persons
consumed food or beverage item were calculated based on the with a high school education, obese persons, and persons trying to
quantity of food and beverages reported and the corresponding lose weight (P , 0.001) (Table 1).
nutrient contents by the National Center for Health Statistics
(NCHS). Given that only the NHANES 1999–2004 included
a second dietary recall, we only use the first dietary recall from Overall adult beverage consumption
each survey for this analysis. The percentage of adults consuming beverages, per capita
We identified 6 mutually exclusive beverage categories in the caloric consumption, and the daily caloric contribution among
NHANES III (from 549 beverage items) and the NHANES 1999– drinkers on a typical day are given in Table 2. Over the period,
2004 (from 523 beverage items) including 1) SSBs (soda, sport the percentage of drinkers and quantity consumed increased the
drinks, fruit drinks and punches, low-calorie drinks, sweetened most for SSBs compared with the other beverage categories. In
tea, and other sweetened beverages), 2) 100% juice, 3) diet 1999–2004, 63% of adults consumed SSBs on a given day, up
beverages, 4) milk (including flavored milk), 5) coffee or tea, from 58% in 1988–1994 (P , 0.001). Over the period, SSBs
and 6) alcohol (23). See Appendix A for more details. Of note, represented the largest source of beverage calories for adults.
some milk, coffee or tea, or alcoholic beverages may have added From 1988–1994 to 1999–2004, daily per capita consumption of
sugar. To relate our results to dietary guidelines and inform SSBs increased by 46 kcal/d (P , 0.001), and average daily
intervention strategies, we used kilocalories (1 kcal ¼ 4.2 kJ) intake among adults who consumed SSBs increased by 55 kcal/d
and fluid ounces (1 oz ¼ 28.57 mL) as 2 primary measures to (P , 0.001). Alcohol was the second largest source of adult per
evaluate consumption patterns. Of note, information on con- capita beverage calories. For other beverage categories, the
sumption location (where the beverage was consumed) was only percentage of milk drinkers declined most over the period (P ¼
available in 1999–2004. 0.045) followed by coffee and tea drinkers (P , 0.001). The per
374 BLEICH ET AL
TABLE 1
Characteristics of US adults (aged 20 y) in the third National Health and Nutrition Examination Survey (NHANES III;
1988–1994) and NHANES 1999–20041
1988–1994 1999–2004 P for trend
Total [n (%)] 15,979 (100) 13,431 (100)
Sex [n (%)]
Male 7470 (47) 6364 (48) 0.443
Female 8509 (53) 7067 (52)
Race-ethnicity [n (%)]
Non-Hispanic white 6654 (83) 6836 (75) ,0.001
Non-Hispanic black 4466 (12) 2585 (11)
Mexican American 4335 (6) 3637 (13)
Age [n (%)]
20–44 y 7885 (57) 5713 (50) ,0.001
45–64 y 3722 (26) 3692 (33)
65 y 3885 (17) 3589 (18)
Education [n (%)]
Less than high school 6428 (25) 4344 (20) ,0.001
High school (or GED) 4874 (34) 3178 (26)
More than high school 4571 (41) 5883 (54)
Employment status [n (%)]

Downloaded from ajcn.nutrition.org by guest on September 5, 2012


Unemployed 6866 (32) 6148 (36) 0.002
Employed 9110 (68) 7278 (64)
Income [n (%)]
Lower income2 6873 (30) 5165 (32) 0.138
Higher income2 7621 (70) 7117 (68)
Body weight status [n (%)]
Normal weight [BMI (kg/m2) 18.5–24.9] 6303 (45) 4100 (34) ,0.001
Overweight (BMI 25–29.9) 5555 (33) 4693 (34)
Obese (BMI  30) 4086 (22) 4269 (31)
Weight-loss intention [n (%)]
Currently trying to lose weight 5856 (41) 3834 (35) ,0.001
Not currently trying to lose weight 10,118 (59) 8250 (65)
1
Percentage of US population estimated with weights to adjust for unequal probability of sampling.
2
Income level was dichotomized based on the poverty:income ratio (ratio of annual family income to federal poverty
line). Lower income refers to persons at or below 130% of poverty, which represents eligibility threshold for the federal
food stamp program.

capita consumption and daily caloric contribution from milk and Per capita intake of SSBs by age, sociodemographics,
coffee or tea remained relatively constant. obesity, and weight-loss intention
Per capita consumption of SSBs and its corresponding percentage
contribution to daily energy intake by age, sociodemographic
Consumption patterns of SSBs and other beverages by sex, groups, weight, and weight-loss intention are shown in Table 4.
race-ethnicity, and age Over the period, per capita consumption of SSBs increased signif-
Overall, the percentage of SSB drinkers increased significantly icantly from 158 kcal/d in 1988–1994 to 203 kcal/d in 1999–2004
from 58% in 1988–1994 to 63% in 1999–2004 (P , 0.001) (P , 0.001) with the biggest increase among young adults. For
(Table 3). For both sexes and all race-ethnicity groups, young adults aged 20–44 y, per capita SSB calories averaged 231 kcal/d
adults (aged 20–44 y) had the highest percentage of SSB (9% total daily intake) in 1988–1994 and 289 kcal/d (12% total
drinkers in 1988–1994 (68%) and in 1999–2004 (72%), a sig- daily intake) in 1999–2004 (P , 0.001). Among young adults, per
nificant upward trend (P ¼ 0.024). The elderly (aged 65 y) had capita consumption of SSBs increased significantly for all de-
the lowest percentage at both time points, but it did increase mographic, body weight, and weight-loss groups. In 1999–2004,
significantly over the period (P , 0.001). Among the race- the caloric contribution of SSBs to overall daily intake was the
ethnicity groups, young black men had the highest percentage of highest among young adults with less than a high school education
SSB drinkers in 1988–1994 (78%) and in 1999–2004 (82%), and and among persons with lower income (14%). Among all race-
the increase over the period was significant (P , 0.05). ethnicity groups, overall per capita consumption was the highest
Notable patterns were also observed for diet drinks, milk, and among blacks. For all ages combined, the increase in per capita
coffee or tea. The percentage of diet drinkers increased significantly SSB consumption was significant for both sexes and all race-
among the elderly overall (13–16%; P ¼ 0.032) and elderly whites ethnicity, education, income, weight status, and intention-to-lose-
(12–16%; P ¼ 0.022). The percentage of milk drinkers declined weight groups (P , 0.05).
most among elderly blacks (57–41%; P , 0.001). Coffee and tea The average quantity and caloric contribution among adults
drinking declined most among young whites (54–48%; P ¼ 0.006) who had 1 consumption occasion of SSBs on the previous day
and young men (51–45%; P ¼ 0.002). are given in Table 5. In 1988–1994, daily consumption of SSBs
INCREASING CONSUMPTION OF SSBs AMONG US ADULTS 375
TABLE 2
Percentage of US adults (aged 20 y) consuming beverages and per capita caloric contribution on the surveyed day,
by beverage, in the third National Health and Nutrition Examination Survey (NHANES III, 1988–1994) and NHANES
1999–20041
1988–1994 1999–2004 Mean difference P for trend
% % %
Consumed beverages on the surveyed day (%)
Had SSBs 58 6 12 63 6 1 5 ,0.001
Had diet 18 6 1 17 6 1 21 0.435
Had juice 20 6 1 20 6 1 0 0.459
Had milk 55 6 1 48 6 1 27 ,0.001
Had coffee or tea 64 6 1 60 6 1 24 0.003
Had alcohol 23 6 2 23 6 2 0 0.748
Per capita caloric contribution (kcal/d)
From SSBs 157 6 5 203 6 5 46 ,0.001
From juice 31 6 1 32 6 1 2 0.420
From milk 92 6 3 84 6 2 29 0.045
From coffee or tea 8 6 0 11 6 1 3 ,0.001
From alcohol 88 6 5 99 6 5 11 0.124
Daily caloric contribution among drinkers (kcal/d)

Downloaded from ajcn.nutrition.org by guest on September 5, 2012


From SSBs 239 6 5 294 6 5 55 ,0.001
From juice 151 6 3 158 6 4 7 0.051
From milk 174 6 4 185 6 3 11 0.030
From coffee or tea 13 6 1 19 6 1 6 ,0.001
From alcohol 376 6 13 418 6 16 42 0.010
1
SSBs, sugar-sweetened beverages. Multivariate regression was used to adjust for race-ethnicity, sex, age, education,
marital status, income, and employment status. SEM , 0.05 is listed as 0.
2
Mean 6 SEM (all such values).

was highest among obese adults (26 oz). In 1999–2004, daily con- to lose weight (313 compared with 278 kcal/d; P , 0.001). Of
sumption of SSBs was highest among men (32 oz). From 1988– note, whereas per capita consumption and the daily caloric con-
1994 to 1999–2004, average daily consumption increased from 22 tribution from SSBs was lower in both survey periods among
to 28 oz (P , 0.001), average consumption rose from 239 to 294 overweight-obese persons trying to lose weight, the consumption
kcal/d (P , 0.001), and the average serving size per SSB con- level increased over time, regardless of weight-loss intention,
sumption occasion increased from 11 oz to 17 oz (P , 0.001). For which is consistent with the trend observed in Table 4.
all groups, average daily consumption and average consumption per Reductions in the percentage of overweight-obese SBB
occasion increased significantly. Not only did more adults drink drinkers among adults trying to lose weight were paralleled by
SSBs but they also drank more each time they consumed SSBs. In increases in the percentage of diet drinkers among adults with
1999–2004, average consumption per drinking occasion was weight-loss intention. In NHANES III and NHANES 1999–2004,
highest among males and whites (19 oz). Between the 2 time peri- the percentage of diet drinkers was significantly higher among
ods, the caloric contribution of SSBs increased the most among men overweight-obese persons trying to lose weight than among
(271–338 kcal/d) and persons with lower income (252–319 kcal/d). persons who were not (P , 0.001).
For other beverage categories, overweight-obese persons with
Beverage consumption among overweight-obese adults by weight-loss intention consumed less alcohol (NHANES III: 92
weight-loss intention compared with 67 kcal/d; P ¼ 0.006) and milk per capita than
persons not trying to lose weight (NHANES III: 97 compared
The beverage consumption patterns among overweight-obese with 83 kcal/d; P ¼ 0.016; NHANES 1999–2004: 88 compared
adults by weight-loss intention are given in Table 6. In both with 74 kcal/d; P ¼ 0.018).
NHANES III and NHANES 1999–2004, the percentage of over-
weight-obese adults who drank SSBs was significantly lower
among persons with weight-loss intention than for persons with- Location of consumption of types of SSBs
out weight-loss intention (P , 0.001). Per capita consumption and Per capitaconsumption of SSBs by consumption location, sex, and
the daily caloric contributions among SSB drinkers followed the age is shown in Figure 1. The largest share of SSB calories were
same pattern. SSBs, on average, contributed more per capita consumed at home, a pattern that increased with age. Approximately
calories to the diet of persons not trying to lose weight than to half of SSB calories were consumed at home by young adults
persons who did try to lose weight (NHANES III: 178 compared compared with approximately three-fourths consumed at home by
with 146 kcal/d; P , 0.001; NHANES 1999–2004: 231 compared the elderly. Among young adults, a sizable amount of SSB calories
with 186 kcal/d; P , 0.001). Likewise, among persons who re- were consumed in restaurants or cafeterias (15%) and at work (20%).
ported 1 consumption occasion of SSBs on the previous day in The relative contribution of each beverage type to per capita
1999–2004, average consumption was significantly higher among SSB consumption by age is shown in Figure 2. Overall, soda
persons without weight-loss intention than among persons trying represented the largest share of SSB calories and comprised
376 BLEICH ET AL
TABLE 3
Percentage of US adults reporting 1 consumption occasion of beverages on the surveyed day, by sex, age, and
race-ethnicity, in the third National Health and Nutrition Examination Survey (NHANES III, 1988–1994) and NHANES
1999–20041
All (age 20 y) Aged 20–44 y Aged 45–64 y Aged 65 y

1988–1994 1999–2004 1988–1994 1999–2004 1988–1994 1999–2004 1988–1994 1999–2004


%
Overall
Had SSBs 58 6 1 63 6 12 68 6 2 72 6 12 50 6 2 57 6 2 36 6 2 43 6 22
Had diet 17 6 1 17 6 0 17 6 1 14 6 12 21 6 2 22 6 2 13 6 1 16 6 22
Had juice 21 6 1 20 6 0 16 6 1 17 6 1 21 6 2 18 6 1 29 6 1 28 6 1
Had milk 55 6 1 48 6 12 52 6 2 45 6 12 56 6 2 47 6 22 69 6 2 63 6 22
Had coffee or tea 62 6 1 60 6 1 50 6 2 45 6 22 75 6 2 72 6 22 80 6 2 81 6 1
Had alcohol 24 6 2 23 6 2 27 6 2 26 6 2 22 6 2 22 6 2 19 6 3 15 6 2
Women
Had SSBs 55 6 1 59 6 12 64 6 2 67 6 2 48 6 2 55 6 22 33 6 2 40 6 22
Had diet 21 6 1 21 6 1 21 6 2 18 6 1 25 6 2 26 6 2 13 6 1 15 6 2
Had juice 21 6 1 20 6 1 16 6 2 17 6 2 22 6 2 19 6 2 31 6 2 27 6 2
Had milk 54 6 1 48 6 12 52 6 2 45 6 22 55 6 2 48 6 22 69 6 2 63 6 32
Had coffee or tea 59 6 2 60 6 2 49 6 2 47 6 2 72 6 2 70 6 2 77 6 3 82 6 22

Downloaded from ajcn.nutrition.org by guest on September 5, 2012


Had alcohol 18 6 2 16 6 2 21 6 2 18 6 2 16 6 3 16 6 2 14 6 3 11 6 2
Men
Had SSBs 62 6 2 65 6 12 74 6 2 76 6 2 54 6 3 59 6 3 39 6 2 47 6 32
Had diet 14 6 1 14 6 1 14 6 2 11 6 12 16 6 1 20 6 22 12 6 1 15 6 2
Had juice 20 6 1 19 6 2 17 6 2 17 6 2 19 6 2 17 6 2 26 6 2 27 6 2
Had milk 55 6 1 49 6 12 51 6 2 46 6 22 55 6 3 47 6 32 69 6 3 63 6 32
Had coffee or tea 64 6 1 61 6 2 51 6 2 45 6 22 76 6 2 73 6 2 83 6 2 81 6 2
Had alcohol 30 6 2 31 6 2 34 6 2 35 6 2 28 6 2 31 6 2 25 6 4 22 6 3
Non-Hispanic whites
Had SSBs 56 6 1 60 6 12 67 6 2 70 6 2 48 6 2 55 6 22 34 6 2 41 6 22
Had diet 20 6 1 19 6 1 20 6 2 16 6 12 23 6 2 25 6 2 12 6 1 16 6 22
Had juice 20 6 1 19 6 1 15 6 2 15 6 2 20 6 2 16 6 12 29 6 2 27 6 1
Had milk 57 6 1 50 6 12 54 6 2 46 6 22 58 6 2 50 6 2 71 6 2 65 6 22
Had coffee or tea 65 6 1 63 6 2 54 6 2 48 6 22 77 6 2 75 6 2 82 6 2 83 6 2
Had alcohol 23 6 2 23 6 2 26 6 2 27 6 2 21 6 2 22 6 2 19 6 3 15 6 2
Non-Hispanic blacks
Had SSBs 73 6 2 76 6 12 78 6 2 82 6 12 68 6 3 71 6 3 58 6 4 63 6 3
Had diet 10 6 1 8 6 1 8 6 2 5 6 1 14 6 2 11 6 2 15 6 3 13 6 3
Had juice 20 6 2 24 6 22 18 6 2 24 6 22 20 6 2 22 6 3 26 6 4 25 6 4
Had milk 38 6 2 32 6 12 33 6 2 31 6 2 42 6 4 28 6 32 57 6 3 41 6 32
Had coffee or tea 41 6 2 38 6 2 30 6 2 25 6 22 61 6 3 51 6 42 64 6 4 65 6 3
Had alcohol 22 6 2 19 6 1 28 6 3 21 6 22 17 6 2 20 6 2 9 6 2 12 6 3
Mexican Americans
Had SSBs 69 6 2 70 6 2 76 6 2 79 6 2 52 6 4 56 6 4 35 6 3 46 6 42
Had diet 14 6 2 13 6 1 11 6 2 10 6 2 22 6 3 18 6 3 19 6 5 11 6 3
Had juice 17 6 1 21 6 22 15 6 2 20 6 22 18 6 3 20 6 3 31 6 4 30 6 4
Had milk 55 6 3 48 6 32 55 6 3 48 6 32 52 6 4 47 6 4 79 6 2 70 6 3
Had coffee or tea 53 6 2 57 6 2 46 6 3 47 6 3 72 6 4 70 6 4 83 6 3 87 6 2
Had alcohol 21 6 1 20 6 1 22 6 2 21 6 2 26 6 4 21 6 3 9 6 2 12 6 4
1
All values are means 6 SEMs. SSBs, sugar-sweetened beverages. Multivariate regression was used to adjust for sex,
race-ethnicity, education, marital status, income, and employment status. SEM , 0.05 is listed as 0.
2
P value for trend , 0.05.

’60% of the total. Fruit drinks (which include all fruit punches DISCUSSION
and fruit nectars with added sugar) were the second largest The 2005 Dietary Guidelines for Americans recommends
source, representing about a quarter of total SSB calories. The drinks without added sugar (24). Yet, our study shows that SSBs,
contribution of soda to overall SSB consumption declined with which provide little nutritional benefit, represent a considerable
age, whereas the contribution of fruit drink increased with age. source of total adult daily intake and is the largest source of
Over the survey period, the relative contribution from both soda beverage calories. In 1999–2004, two-thirds of adults (63%)
and fruit drink increased across all age groups. For example, drank SSBs, averaging 28 oz/d, 17 oz per consumption occasion,
soda contributed 61% of all SSB calories in 1999–2004, which and 293 calories daily (15% of recommended 2000 kcal/d diet).
was up from 56% in 1988–94. More adults are drinking SSBs (primarily soda; ’60%), and,
INCREASING CONSUMPTION OF SSBs AMONG US ADULTS 377
TABLE 4
US adult per capita consumption of sugar-sweetened beverages (SSBs), measured in caloric contributions, in the third National Health and Nutrition
Examination Survey (NHANES III, 1988–1994) and NHANES 1999–20041
All (age 20 y) Age 20–44 y Age 45–64 y Age 65 y

1988–1994 1999–2004 1988–1994 1999–2004 1988–1994 1999–2004 1988–1994 1999–2004


kcal/d
Overall 158 6 5 (7) 203 6 5 (9)2 231 6 7 (10) 289 6 7 (12)2 124 6 6 (6) 160 6 6 (8) 68 6 5 (4) 83 6 5 (5)2
Sex
Female 128 6 4 (7) 163 6 5 (8)2 170 6 11 (10) 214 6 11 (12)2 104 6 6 (6) 136 6 7 (7)2 62 6 9 (4) 71 6 7 (5)
Male 186 6 7 (7) 243 6 7 (9)2 276 6 12 (10) 341 6 12 (13)2 146 6 6 (6) 183 6 9 (8)2 72 6 7 (4) 96 6 6 (5)2
Race-ethnicity
Non-Hispanic white 160 6 6 (7) 205 6 6 (9)2 251 6 10 (11) 307 6 11 (13)2 113 6 7 (5) 150 6 8 (7)2 56 6 5 (3) 72 6 5 (4)2
2 2 2
Non-Hispanic black 175 6 4 (9) 234 6 8 (11) 235 6 8 (10) 308 6 12 (13) 161 6 6 (9) 205 6 9 (11) 107 6 10 (7) 125 6 11 (8)
Mexican American 153 6 7 (7) 192 6 8 (9)2 225 6 8 (10) 274 6 13 (11)2 108 6 11 (5) 128 6 9 (7) 65 6 10 (4) 86 6 9 (5)
Education
Less than high school 174 6 7 (8) 223 6 6 (10)2 249 6 16 (11) 325 6 12 (14)2 133 6 12 (7) 164 6 9 (9)2 71 6 9 (4) 91 6 7 (6)2
High school or GED 169 6 6 (8) 220 6 8 (9)2 250 6 10 (11) 318 6 11 (13)2 134 6 10 (6) 171 6 12 (8)2 68 6 8 (4) 84 6 9 (5)2
More than high school 155 6 8 (7) 194 6 8 (8)2 218 6 10 (9) 263 6 11 (11)2 117 6 12 (5) 152 6 9 (7)2 73 6 10 (4) 84 6 8 (4)
Income3
Lower income 168 6 7 (8) 229 6 7 (10)2 256 6 8 (11) 334 6 11 (14)2 130 6 12 (7) 184 6 11 (9)2 67 6 14 (4) 84 6 10 (5)

Downloaded from ajcn.nutrition.org by guest on September 5, 2012


Higher income 156 6 6 (7) 195 6 5 (9)2 232 6 10 (10) 279 6 8 (11)2 119 6 7 (6) 150 6 7 (7) 71 6 6 (4) 85 6 7 (5)2
Body weight status4
Normal weight 153 6 7 (7) 198 6 7 (9)2 222 6 9 (10) 275 6 11 (11)2 116 6 11 (6) 160 6 13 (8)2 69 6 7 (4) 79 6 8 (5)
Overweight 155 6 6 (7) 194 6 6 (9)2 232 6 10 (10) 277 6 12 (12)2 118 6 9 (6) 152 6 9 (8)2 59 6 7 (4) 80 6 6 (5)2
Obese 166 6 5 (8) 220 6 8 (10)2 244 6 12 (10) 323 6 14 (13)2 126 6 7 (6) 157 6 8 (8)2 75 6 13 (4) 94 6 13 (5)2
Weight-loss intention5
Not trying to lose weight 162 6 6 (7) 215 6 6 (9)2 234 6 9 (10) 303 6 10 (12)2 129 6 8 (7) 175 6 9 (9)2 69 6 6 (4) 86 6 6 (5)2
Trying to lose weight 151 6 6 (7) 190 6 7 (9)2 218 6 11 (10) 266 6 11 (12)2 116 6 6 (6) 143 6 8 (7) 2 68 6 10 (4) 85 6 11 (5)2
1
All values are means 6 SEMs; percentage of contribution to daily energy intake in parentheses. GED, General Education Development test that certifies
high school-level skills. Multivariate regression was used to adjust for sex, race-ethnicity, education, marital status, income, and employment status. SEM ,
0.05 is listed as 0.
2
P for trend , 0.05.
3
Income level was dichotomized based on the poverty:income ratio and eligibility for food assistance programs (ie, 130% of the poverty level).
4
Normal weight was defined as a BMI (in kg/m2) from 18.5 to 24.9, overweight as BMI from 25 to 29.9, and obese as BMI  30.
5
Categories of trying to lose weight include all BMI groups.

among SSB drinkers, average caloric consumption and quantity percentage of SSB drinkers (61% compared with 67%, 1999–
consumed had increased, changes that parallel the rising prev- 2004) and the amount of SSBs consumed (186 compared with
alence of adult obesity (2) and type 2 diabetes (29). 231 kcal/d, 1999–2004) was lower among overweight-obese
An additional key finding is the considerable variation in adult persons with weight-loss intention compared with persons with-
SSB consumption by sociodemographic characteristics, with the out this intention. This suggests that persons trying to lose weight
highest rates concentrated among populations at highest risk of are drinking less SSBs and selecting SSBs with lower energy
obesity (2, 3). Overall, SSB consumption declined steeply with content. Yet, per capita consumption of SSBs increased among all
age, which may be partially because of a cohort effect (ie, weight groups over the period regardless of weight-loss intention.
younger generations have increased their SSB consumption) However, differences between the groups are small, which in-
(30). In 1999–2004, young adults had the highest prevalence of dicates that greater efforts are needed to empower persons trying
SSB drinkers (72%) and were the highest consumers of SSB to lose weight to adopt healthy eating behaviors.
calories (289 kcal/d) compared with older age groups. Among Both similarities and differences are found between our findings
race-ethnicity groups, blacks had the highest percentage of SSB and previous research. This study is consistent with previous work
drinkers (76%) and the highest per capita consumption of SSBs that identified a consistent upward trend in calories from SSBs (5, 7,
(234 kcal/d) in 1999–2004, followed by Mexican Americans. 9–11, 32) and high SSB consumption at home (5). Compared with
Among all young adults, the caloric contribution of SSBs was trends in US children and adolescents, in whom the percentage of
highest among persons with lower income and persons with less SSB drinkers remained constant from 1988–1994 to 1999–2004
than a high school education. and in whom per capita caloric consumption of SSB increased
Increasingly, US adults are reporting a desire to lose weight (30), we observed increases in both the percentage of SSB drinkers
(31). Yet, the extent to which this maps to changes in beverage and the quantify of SSBs consumed among adults. This study
consumption is not well known. This study offers useful in- improves our understanding of adult beverage consumption by
formation about SSB consumption patterns by weight-loss in- focusing on variations by sociodemographic characteristics,
tention. Our results indicate that overweight-obese adults trying weight status, and intention to lose weight. Previous studies have
to lose weight (compared with those without weight-loss in- largely focused on overall averages, ignoring the heterogeneity by
tention) are moving in the right direction. In both surveys, the subpopulation. The considerable gradient we observe by age, for
378 BLEICH ET AL
TABLE 5
Average quantity and caloric contributions from sugar-sweetened beverages among US adults who had 1 consumption
occasion on the surveyed day, in the third National Health and Nutrition Examination Survey (NHANES III, 1988–1994)
and NHANES 1999–20041
Volume per consumption
Daily consumption occasion Energy intake

1988–1994 1999–2004 1988–1994 1999–2004 1988–1994 1999–2004


oz oz kcal
Overall 22 6 1 28 6 12 11 6 0 17 6 02 239 6 6 294 6 52
Sex
Female 19 6 1 24 6 12 10 6 0 15 6 02 207 6 6 249 6 62
Male 25 6 1 32 6 12 13 6 0 19 6 02 271 6 10 338 6 82
Race-ethnicity
Non-Hispanic white 25 6 1 31 6 12 13 6 0 19 6 12 262 6 8 316 6 82
Non-Hispanic black 22 6 0 29 6 12 12 6 0 17 6 02 247 6 4 311 6 72
Mexican American 21 6 1 26 6 12 12 6 0 15 6 12 233 6 8 275 6 92
Education
Less than high school 24 6 1 29 6 12 12 6 0 18 6 02 253 6 9 310 6 82
High school or GED 23 6 1 30 6 12 12 6 0 17 6 02 249 6 9 302 6 102
More than high school 22 6 1 28 6 12 12 6 1 17 6 02 239 6 11 293 6 112

Downloaded from ajcn.nutrition.org by guest on September 5, 2012


Income3
Lower income 23 6 1 30 6 12 12 6 0 18 6 12 252 6 8 319 6 92
Higher income 22 6 1 28 6 12 12 6 0 17 6 02 235 6 8 283 6 62
Body weight status4
Normal weight 21 6 1 27 6 12 11 6 0 16 6 12 228 6 9 289 6 82
Overweight 22 6 1 27 6 12 12 6 0 17 6 02 236 6 7 276 6 72
Obese 26 6 1 31 6 12 13 6 0 18 6 12 270 6 7 317 6 92
Weight-loss intention5
Not trying to lose weight 22 6 1 29 6 12 11 6 0 17 6 02 238 6 7 301 6 72
Trying to lose weight 23 6 1 28 6 12 12 6 0 17 6 02 243 6 8 285 6 92
1
All values are mean 6 SEM. GED, General Education Development test that certifies high school-level skills.
Conversion factor (1 oz ¼ 28.57 mL). Multivariate regression was used to adjust for age, sex, race-ethnicity, education,
marital status, income, and employment status. SEM , 0.05 is listed as 0.
2
P for trend , 0.05.
3
Income level was dichotomized based on the poverty:income ratio and eligibility for food assistance programs
(ie, 130% of the poverty level).
4
Normal weight was defined as a BMI (in kg/m2) from 18.5 to 24.9, overweight as BMI from 25 to 29.9, and obese as
BMI  30.
5
Categories of trying to lose weight include all BMI groups.

example, suggests that prior research focused on the entire adult trend toward decreased SSB consumption among overweight-
population may have underestimated the energy imbalance among obese persons trying to lose weight, an easy and concrete, be-
younger adults. havioral target that may be best encouraged by physicians and
Our focus on multiple subpopulations makes it possible to health care professionals.
highlight several potential opportunities to reduce overall caloric In general, strategies that promote reductions in SSB intake
intake as well as to identify concerning patterns for future study. may be easier to attain than strategies that promote increases in
To reduce adult intake of SSBs, the considerable consumption of physical activity, given the time trade-off between energy intake
SSBs in the workplace, particularly among young adults, sug- and expenditure. For example, an 85-kg man (187.4 lb) would
gests that initiatives focused in this area may have a significant need to walk for 50 min (rather than sit) to burn off 1 can (12 oz)
effect. To achieve this goal, one possibility might be to reduce the of soda (140 kcal/d).
standard serving size of SSBs in the workplace (eg, from a 20-oz Additional research is needed to understand the drivers of
bottle to a 12-oz can), given our finding of increased consumption increased SSB consumption among subgroups with higher in-
per drinking occasion. Another might be to reduce the ready take. For example, the recent Institute of Medicine report that
availability of SSBs, which undermines initiatives aimed at re- focused on food marketing to children (35) found that ethnic
ducing consumption, or to replace SSB beverage options with minorities living in poorer neighborhoods have fewer healthier
noncaloric alternatives. This latter strategy has been effective in options (36, 37). From a policy perspective, targeted programs
adolescents (33). Efforts to reduce SSB consumption in the among these subpopulations, promoting reduced consumption of
workplace, where we found that young adults consume a fifth of empty calories, may be useful for reducing the upward trend in
their SSB calories, would complement the growing interest consuming SSBs.
among employers of investing in health-promotion strategies The present study has several limitations. First, our reliance on
(34). An additional approach might be to encourage the current single 24-h dietary recalls may introduce inaccuracy and bias to
INCREASING CONSUMPTION OF SSBs AMONG US ADULTS 379
TABLE 6
Percentage of overweight-obese US adults (aged 20 y) consuming beverages and per capita caloric contribution on the
survey day, by weight-loss intention and beverage, in the third National Health and Nutrition Examination Survey
(NHANES III, 1988–1994; n ¼ 9639) and NHANES 1999–2004 (n ¼ 7862)1
1988–1994 1999–2004

Not trying to lose Trying to lose Not trying to lose Trying to lose
weight weight weight weight
Consumed beverages on the
surveyed day (%)
Had SSBs 62 6 2 54 6 22 67 6 2 61 6 22
Had diet 17 6 2 26 6 22 15 6 1 21 6 12
Had juice 19 6 1 17 6 1 20 6 1 19 6 1
Had milk 54 6 2 54 6 2 47 6 0 46 6 0
Had coffee or tea 63 6 2 62 6 2 59 6 0 59 6 0
Had alcohol 22 6 2 18 6 22 18 6 0 20 6 0
Per capita caloric contribution
(kcal/d)
From SSBs 178 6 6 146 6 52 231 6 6 186 6 72
From juice 29 6 2 29 6 1 34 6 2 29 6 22
From milk 97 6 5 83 6 42 88 6 4 74 6 42

Downloaded from ajcn.nutrition.org by guest on September 5, 2012


From coffee or tea 9 6 0 7 6 02 10 6 1 11 6 1
From alcohol 92 6 5 67 6 62 89 6 6 91 6 8
Daily caloric contribution among
drinkers (kcal/d)
From SSBs 261 6 8 240 6 7 313 6 7 278 6 92
From juice 146 6 6 150 6 7 161 6 6 148 6 6
From milk 184 6 7 161 6 62 195 6 7 171 6 6
From coffee or tea 15 6 1 12 6 12 19 6 1 21 6 2
From alcohol 359 6 21 336 6 20 426 6 19 409 6 33
1
All values are means 6 SEMs. Overweight-obese was defined as BMI (in kg/m2)  25. SEM , 0.05 is listed as 0.
SSBs, sugar-sweetened beverages. Multivariate regression was used to adjust for age, sex, race-ethnicity, education, marital
status, income, and employment status.
2
P for between-group difference , 0.05.

our analyses because of underreporting, unreliability, and con- persons not trying to lose weight may be smaller than our results
version error. Previous research indicates that adults underreport indicate. However, available evidence does suggest better recall
their dietary consumption by ’25% (38, 39). In addition, re- accuracy with packaged beverage items such as SSBs (41). A
strained eating has been associated with underreporting of ca- single 24-h dietary recall may not accurately represent usual
loric intake (40). Therefore, the true difference in SSB dietary intake for a person. Lack of reliability of the dietary
consumption between persons with weight-loss intention and recall for overall eating habits will reduce the precision of our

FIGURE 1. Per capita calories from sugar-sweetened


beverages by location of consumption according to
the National Health and Nutrition Examination
Survey (NHANES) 1999–2000. Note: Multivariate
regression was used to adjust for sex, race-ethnicity,
education, marital status, income, and employment
status.
380 BLEICH ET AL

FIGURE 2. Per capita calories from sugar-sweetened


beverages by beverage type, third National Health
and Nutrition Examination Survey (NHANES III)
compared with NHANES 1999–2004. Note:
Multivariate regression was used to adjust for age,
sex, race-ethnicity, education, marital status, income,
and employment status.

Downloaded from ajcn.nutrition.org by guest on September 5, 2012


estimates, but it will not bias our regression estimates when total The authors’ responsibilities were as follows—SNB, YCW, and SLG: con-
energy intake is the dependent variable (42). Inaccuracy exists in ceived the study and developed the hypotheses; SNB (guarantor) analyzed the
converting reported beverage consumption to energy intake data and drafted the manuscript; SNB, YCW, YW, and SLG: contributed to
the interpretation of study findings and to the final draft. None of the authors
because the assumptions on serving size and food composition
had a personal or financial conflict of interest.
are defined by the food and nutrient database. This standard
database assumes a representative nutritional content for a given
food or beverage. The inevitable variation in actual intake and REFERENCES
reporting bias may introduce measurement errors, particularly 1. Malik VS, Schulze MB, Hu FB. Intake of sugar-sweetened beverages
for the estimation of total energy intake. However, this error is and weight gain: a systematic review. Am J Clin Nutr 2006;84:274–88.
likely less significant for packaged, standard-sized beverages. 2. Ogden CL, Carroll MD, Curtin LR, McDowell MA, Tabak CJ, Flegal
KM. Prevalence of overweight and obesity in the United States,
Second, although we used multivariate regression models to 1999–2004. JAMA 2006;295:1549–55.
adjust for demographic variables, our inferences on secular 3. Wang Y, Beydoun MA. The obesity epidemic in the United States—gender,
changes and beverage consumption patterns between the 2 sur- age, socioeconomic, racial/ethnic, and geographic characteristics: a sys-
veys may remain constrained by changes in other demographic tematic review and meta-regression analysis. Epidemiol Rev 2007;29:
6–28.
variables, but the 2 samples had comparable distributions of sex, 4. Schulze MB, Manson JE, Ludwig DS, et al. Sugar-sweetened beverages,
age, employment status, and income status. Third, the NHANES weight gain, and incidence of type-2 diabetes in young and middle-aged
data are cross-sectional, which only allows us to address asso- women. JAMA 2004;292:927–34.
ciations rather than causality. Fourth, the intention-to-lose- 5. Nielsen SJ, Popkin BM. Changes in beverage intake between 1977 and
2001. Am J Prev Med 2004;27:205–10.
weight question asks respondents about the previous 12 mo, 6. Flegal KM, Campbell SM, Johnson CL. Prevalence and trends in obesity
whereas the 24-h recall focuses on 1 d. Temporal differences may among US adults, 1999–2000. JAMA 2002;288:1723–7.
limit our ability to fully capture eating practices of persons trying 7. Duffey KJ, Popkin BM. Shifts in patterns and consumption of beverages
to lose weight. In addition, differences in survey methods between between 1965 and 2002. Obesity (Silver Spring) 2007;15:2739–47.
8. Huffman L, West DS. Readiness to change sugar sweetened beverage
the 2 study periods (eg, NHANES 1999–2004 used computer-
intake among college students. Eat Behav 2007;8:10–4.
assisted interviewing for the dietary interview, but NHANES III 9. Popkin BM, Nielsen SJ. The sweetening of the world’s diet. Obes Res
did not) may affect the comparability of data overtime. 2003;11:1325–32.
To conclude, SSB consumption has increased dramatically in 10. Nielsen SJ, Siega-Riz AM, Popkin BM. Trends in energy intake in U.S.
the past decade in the United States, in parallel with the rising between 1977 and 1996: similar shifts seen across age groups. Obes Res
2002;10:370–8.
prevalence of obesity and type 2 diabetes. Large epidemiologic 11. Storey ML, Forshee RA, Anderson PA. Beverage consumption in the US
studies provide strong evidence for the independent effect of population. J Am Diet Assoc 2006;106:1992–2000.
SSBs on weight gain and type 2 diabetes (1, 4). Efforts to en- 12. McLaren L. Socioeconomic status and obesity. Epidemiol Rev 2007;29:
courage replacing SSBs with low-caloric or noncaloric alter- 29–48.
13. Zhang Q, Wang Y. Socioeconomic inequality of obesity in the United
natives may be an important strategy to reduce consumption of States: do gender, age, and ethnicity matter? Soc Sci Med 2004;58:
empty calories. Physicians and public health professionals are 1171–80.
well positioned to identify and promote concrete behavioral 14. Wang Y, Zhang Q. Are American children and adolescents of low socio-
targets aimed at decreasing adult SSB consumption, making economic status at increased risk of obesity? Changes in the association
between overweight and family income between 1971 and 2002. Am J
awareness of these changes critical among that group. The Clin Nutr 2006;84:707–16.
workplace and home offer key areas of intervention for reducing 15. Horm J, Anderson K. Who in America is trying to lose weight? Ann
the energy imbalance in young adults. Intern Med 1993;119:672–6.
INCREASING CONSUMPTION OF SSBs AMONG US ADULTS 381
16. Levy AS, Heaton AW. Weight control practices of U.S. adults trying to 38. Bingham SA, Gill C, Welch A, et al. Comparison of dietary assessment
lose weight. Ann Intern Med 1993;119:661–6. methods in nutritional epidemiology: weighed records v. 24 h recalls,
17. Serdula MK, Mokdad AH, Williamson DF, Galuska DA, Mendlein JM, food-frequency questionnaires and estimated-diet records. Br J Nutr
Heath GW. Prevalence of attempting weight loss and strategies for 1994;72:619–43.
controlling weight. JAMA 1999;282:1353–8. 39. Briefel RR, Sempos CT, McDowell MA, Chien S, Alaimo K. Dietary
18. Goldstein DJ. Beneficial health effects of modest weight loss. Int J Obes methods research in the third National Health and Nutrition Examina-
Relat Metab Disord 1992;16:397–415. tion Survey: underreporting of energy intake. Am J Clin Nutr 1997;
19. Mattes RD. Dietary compensation by humans for supplemental energy 65(suppl):1203S–9S.
provided as ethanol or carbohydrate in fluids. Physiol Behav 1996;59: 40. Macdiarmid J, Blundell J. Assessing dietary intake: who, what and why
179–87. of under-reporting. Nutr Res Rev 1998;11:231–53.
20. De Castro JM. The effects of the spontaneous ingestion of particular 41. Guthrie HA. Selection and quantification of typical food portions by
foods or beverages on the meal pattern and overall nutrient intake of young adults. J Am Diet Assoc 1984;84:1440–4.
humans. Physiol Behav 1993;53:1133–44. 42. Fox J. Applied regression analysis, linear models and related methods.
21. Jaquier E, Tappy L. Regulation of body weight in humans. Physiol Rev Thousand Oaks, CA: Sage Publications, 1997.
1999;79:451–80.
22. Bleich S, Cutler D, Murray C, Adams A. Why is the developed world
obese? Annu Rev Public Health 2008;29:273–95.
23. Popkin BM, Armstrong LE, Bray GM, Caballero B, Frei B, Willett WC.
A new proposed guidance system for beverage consumption in the
APPENDIX A
United States. Am J Clin Nutr 2006;83:529–42.
24. Department of Health and Human Services and US Department of
Agriculture. Dietary guidelines for Americans. Washington, DC:
Coding definition for nonalcoholic beverage categories:
Department of Health and Human Services, US Department of Agri- NHANES III and NHANES 1999–2004
culture, 2005. 1) Sugar-sweetened beverages (SSBs) includes all sodas, fruit
25. Berkey CS, Rockett HR, Field AE, Gillman MW, Colditz GA. Sugar-added

Downloaded from ajcn.nutrition.org by guest on September 5, 2012


drinks, sport drinks, low-calorie drinks, and other beverages
beverages and adolescent weight change. Obes Res 2004;12:778–88.
26. James J, Thomas P, Cavan D, Kerr D. Preventing childhood obesity by
[sweetened tea, rice drinks, bean beverages, sugar cane bev-
reducing consumption of carbonated drinks: cluster randomised con- erages, horchata (a traditional Mexican beverage made with
trolled trial. BMJ 2004;328:1237–42. rice), and nonalcoholic wines or malt beverages].
27. Ludwig DS, Peterson KE, Gortmaker SL. Relation between consump- a) Sport drink includes all drinks labeled Gatorade or thirst
tion of sugar-sweetened drinks and childhood obesity: a prospective, quencher (3 items in NHANES III, 7 items in NHANES
observational analysis. Lancet 2001;357:505–8. 1999–2004).
28. World Health Organization. Obesity: preventing and managing the b) Fruit drink includes all fruit drinks, fruit juices, and fruit
global epidemic—report of a WHO consultation on obesity. Geneva, nectars with added sugar (166 items in NHANES III, 137
Switzerland: World Health Organization, 1988. items in NHANES 1999–2004).
29. Mokdad AH, Ford ES, Bowman BA, et al. Prevalence of obesity, diabetes,
and obesity-related health risk factors, 2001. JAMA 2003;289:76–9. c) Soda includes all carbonated beverages with added sugar (25
30. Wang CY, Bleich SN, Gortmaker SL. Increasing caloric contribution items in NHANES III, 21 items in NHANES 1999–2004).
from sugar-sweetened beverages and 100% fruit juices among US d) Low-caloric SSBs include all beverages described as ‘‘low-
children and adolescents, 1988–2004. Pediatrics 2008;121:e1604–14. calorie.’’ This includes fruit juices, teas, and fruit drinks (26
31. The Gallup Poll: public opinion 2006. Farmington Hills, MI: Rowman & items in NHANES III, 26 items in NHANES 1999–2004).
Littlefield Publishers Inc, 2005. e) Other SSBs include sweetened tea, rice drinks, bean bev-
32. Wang YC, Gortmaker SL, Sobol AM, Kuntz KM. Estimating the energy erages, sugar cane beverages, horchata, nonalcoholic wines
gap among US children: a counterfactual approach. Pediatrics 2006;118: or malt beverages, etc (40 items in NHANES III, 37 items
e1721–33.
in NHANES 1999–2004).
33. Ebbeling CB, Feldman HA, Osganian SK, Chomitz VR, Ellenbogen SJ,
Ludwig DS. Effects of decreasing sugar-sweetened beverage con-
2) Diet beverages include all diet sodas and sugar-free carbonated
sumption on body weight in adolescents: a randomized, controlled pilot soda water (17 items in NHANES III, 17 items in NHANES
study. Pediatrics 2006;117:673–80. 1999–2004).
34. Partnership to Flight Chronic Disease. Keeping Americans healthy: a 3) Milk includes all whole, low-fat, skim milk, and flavored milk
catalog of successful programs. Available from: http://promisingpractices. (62 items in NHANES III; 55 items in NHANES 1999–2004).
fightchronicdisease.org/uploads/Best_Practice_Catalog.pdf (cited 11 June Flavored milk includes fruit-flavored milk and chocolate milk
2008). (60 items in NHANES III; 69 items in NHANES 1999–2004).
35. Institute of Medicine. Food marketing to children and youth: threat or 4) 100% Juice (FJ) includes all 100% juices (eg, apple and or-
opportunity? Washington, DC: The National Academies Press, 2005. ange) and all unsweetened juices (27 items in NHANES III; 17
36. Powell LM, Auld MC, Chaloupka FJ, O’Malley PM, Johnston LD.
Associations between access to food stores and adolescent body mass
items in NHANES 1999–2004).
index. Am J Prev Med 2007;33(suppl):S301–7. 5) Coffee or tea includes all coffee drinks and unsweetened teas
37. Powell LM, Chaloupka FJ, Bao Y. The availability of fast-food and (59 items in NHANES III; 67 items in NHANES 1999–2004).
full-service restaurants in the United States: associations with neigh- 6) Alcohol includes all alcoholic beverages (64 items in
borhood characteristics. Am J Prev Med 2007;33(suppl):S240–5. NHANES III; 70 items in NHANES 1999–2004).